Bots Dominate the Web: Over 57% of Traffic Is No Longer Human

The Tipping Point: Bots Take Control

For decades, the internet was an ecosystem dominated by humans: people browsing, reading, writing, and sharing. That balance has been broken. According to Cloudflare Radar data, bots now account for between 57% and 58% of HTTP requests for HTML content, compared to 42-43% for humans. Imperva, in its 2025 bad bot report based on data from that year, puts the figure at around 53% for bots, remaining above the 47% human share for the second consecutive year. This phenomenon is not sudden: as early as the 2010s, search engine crawlers and commercial scrapers represented a significant portion of traffic, but the current scale is unprecedented. Generative AI has accelerated the process: bots now not only extract data but also generate content, flooding platforms with synthetic texts.

It's not just about traffic. Generative AI is rewriting the rules of content creation. AI detection company Pangram analyzed platforms like LinkedIn, Medium, Twitter (X), and Reddit, and found that roughly one in four long-form articles was entirely AI-generated. LinkedIn leads the saturation: over 40% of long posts were flagged as fully AI-generated. On X/Twitter, nearly half of articles (46.8%) were either fully or partially synthetic (23.9% fully AI, 22.9% mixed), leaving only 53.2% as fully human. These figures reflect a structural shift: the web is no longer a predominantly human space, but a hybrid ecosystem where machines compete with people on equal footing.

Why It Matters: The Web Becomes a Mirror of Itself

This change has profound implications. The web that humans built is being consumed and replicated by machines, creating a feedback loop where AI trains on AI-generated content. The result is a degradation of quality, originality, and trust. As The Register columnist notes, AI is nothing more than a copy-paste system on an industrial scale. When a language model responds about Linux, for example, it doesn't just pull from authoritative sources like the kernel mailing list, but also from obscure forums and erratic posts. This creates a 'mirror copy' effect: AI replicates errors and biases, amplifying them indiscriminately.

The problem of veracity worsens. AI sounds convincing but often gets things wrong, and its apparent confidence leads many to accept its answers as absolute truths. This is already affecting sectors like journalism, education, and research. For instance, in 2023, a lawyer used ChatGPT to prepare a court brief and cited fictitious cases, leading to sanctions. In education, teachers report an increase in AI-generated assignments containing conceptual errors. Trust in online information erodes, and with it, the foundation of shared knowledge.

Consequences: Who Wins and Who Loses?

For human content creators, the competition is unfair. A non-fiction writer who generated a novel with AI as a joke still earns thousands of dollars in passive income each month, surpassing many professional novelists. This case, reported by The Register, illustrates how AI can economically displace humans even in creative tasks. Meanwhile, copyright lawsuits multiply. The New York Times and other publishers have sued OpenAI, alleging the company deleted training records despite court orders to preserve them. In 2024, a group of visual artists sued Stability AI for using their works without permission. These legal battles will define the future of intellectual property in the age of AI.

For users, the web experience deteriorates. Platforms like LinkedIn and X become filled with generic, repetitive posts, making it hard to find authentic content. Trust in online information erodes. A 2024 study from Stanford University showed that users are unable to distinguish between real and AI-generated news in 40% of cases. This has serious consequences for democracy and informed decision-making. Additionally, companies that rely on web traffic, such as news outlets, see their advertising revenue decline when bots inflate metrics without generating real conversions.

What Readers Should Know

• Not everything you read online is written by humans. Verify sources and be wary of overly perfect or generic texts.
• AI detection tools, like Pangram, can help but are not infallible. A 2024 study found these tools have a 15-20% error rate on short texts.
• Support human creators: subscribe to independent media, share original content, and demand transparency about AI use. Platforms like Medium already label AI-generated content, but implementation is uneven.

“AI is not intelligent at all. It's just a copy-paste of words that are likely to go together. It can sound correct, but often it isn't.” — The Register

The dominance of bots is not inevitable. Governments and platforms can implement regulations and technologies to label AI-generated content and prioritize human-made material. The European Union, with its AI Act, already requires transparency in AI systems. But responsibility also falls on us, the users, to keep the authentic web alive. The history of the internet teaches us that every technological change brings a reaction: from spam to filters, from fake news to verification. Now, facing the tide of AI, the response must be collective and conscious.